It is our hypothesis that viral tests on CSF will facilitate early diagnosis of primary CNS lymphoma (PCNSL) in AIDS patients. This idea follows from (i) our observation that 17 of 17 AIDS PCNSL are positive for expression of the EBER1 EBV gene as detected by in situ hybridization, (ii) the characteristic periventricular location of PCNSL and tendency to meningeal involvement and (iii) preliminary evidence that EBV can be directly detected much more frequently in the CSF from AIDS patients with PCNSL than from AIDS patients without PCNSL. We intend to assess CSF from SIDS patients for the presence of EBV-infected lymphocytes by in situ hybridization, EBV DNA by PCR amplification and antiviral antibodies by serology studies. As information about the predictive value of these studies is accrued, we intend to incorporate this information into a pilot trial of early diagnosis and treatment of PCNSL in AIDS patients. Patients with leptomeningeal involvement will have the opportunity to participate in a trial involving a new mode of administration of intrathecal chemotherapy (an indwelling lumbar ventricular reservoir) . Those with refractory or relapsing leptomeningeal disease will also be eligible for a Phase 1 trial of a new agent (4HC) for intrathecal therapy. We will assess survival, Causes of death, quality of life as measured by neurologic and psychometric testing, and better define the virologic correlates of PCNSL. Material will be collected and banked such that it will be available for parallel study at other institutions to evaluate the exportability of the diagnostic techniques used. This effort will involve a multidisciplinary collaboration of clinical and laboratory investigators drawing upon the resources of well-established programs. If these approaches prove fruitful then the presence of EBV will be important not only in terms of insights regarding etiology and pathogenesis, but as a tumor specific marker which will aid in diagnosis find management of AIDS PCNSL.